From a814d3a7893eb692dafeb0a5ad785af7547718e3 Mon Sep 17 00:00:00 2001 From: ydshieh Date: Thu, 15 Dec 2022 18:45:28 +0100 Subject: [PATCH] Recompile apex in DeepSpeed CI image --- docker/transformers-pytorch-deepspeed-latest-gpu/Dockerfile | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/docker/transformers-pytorch-deepspeed-latest-gpu/Dockerfile b/docker/transformers-pytorch-deepspeed-latest-gpu/Dockerfile index 0b3e9d67375f1e..7234d6881ec131 100644 --- a/docker/transformers-pytorch-deepspeed-latest-gpu/Dockerfile +++ b/docker/transformers-pytorch-deepspeed-latest-gpu/Dockerfile @@ -24,6 +24,11 @@ RUN python3 -m pip install --no-cache-dir ./transformers[deepspeed-testing] RUN python3 -m pip install torch-tensorrt==1.3.0 --find-links https://github.com/pytorch/TensorRT/releases/expanded_assets/v1.3.0 +# recompile apex +RUN python3 -m pip uninstall -y apex +RUN git clone https://github.com/NVIDIA/apex +RUN cd apex && python3 -m pip install --global-option="--cpp_ext" --global-option="--cuda_ext" --no-cache -v --disable-pip-version-check . + # Pre-build **latest** DeepSpeed, so it would be ready for testing (otherwise, the 1st deepspeed test will timeout) RUN python3 -m pip uninstall -y deepspeed # This has to be run (again) inside the GPU VMs running the tests.